<p>An evaluation of site quality was conducted on the Hofmann Forest located in Jones and Onslow Counties of North Carolina. Site quality is a key element in the plan for management to reach its objectives. Accurately measuring site quality can determine the optimal locations for intensively managed plantations and the areas that may be better managed for other purposes. The identification of these areas will allow management to maximize the returns from the forest regardless of the objectives. There are a number of methods available for measuring site quality that includes both direct and indirect methods. The standard method among foresters in the United States is the direct method of the site index approach. The most common indirect method is the soil-site study. The objectives of this research are to examine these two procedures for their reliability and accuracy in determining the site quality on the Hofmann Forest.The foundation that the site index approach is built upon is the theory that tree height growth in relation to age is very sensitive to site quality but is independent of various factors; i.e. stocking density, species compositions, etc (Avery and Burkhart, 1994). Therefore if height growth is not independent of these factors then the site index approach will provide erroneous predictions. The current measure of site quality on the Hofmann Forest was provided by a soil-site study performed by T. S. Coile in 1966. A recent timber inventory of the forest provides a means of testing the accuracy of these measures. An intensive database was constructed from the timber inventory collected by F & W Services, Inc., the soil mapping performed by T. S. Coile, and previous climatic and environmental data for the forest. To achieve the objectives of this research the data was used to develop a non-linear model containing three categories or groups of variables. These categories include variables representing management decisions, soil characteristics, and environmental factors. The model produced an R-squared value of 0.62 with a p-value of <0.0001. A correlation analysis of the height predictions and the actual heights collected in the timber inventory produced a correlation coefficient of 0.79. Nested structural tests found management decisions very significant in influencing height growth. Soil characteristics and precipitation were not significant individually but were when used in conjunction with each other. A correlation analysis was performed on the model predictions and the previous predictions found in Coile's soil-site study, which produced a correlation coefficient of 0.06. Also, the correlation of Coile's predictions and the actual tree heights was 0.22. A forest wide average height at a base age of 25 years was predicted at 10.94 feet lower than that found by Coile. <P>

<p>Handest, Joshua A. Effects of Nutrient Amendments and Genotype on Stand Productivity andCrown Characteristics in Loblolly Pine (Pinus taeda L.) (Under the direction of H. Lee Allen andSteven E. McKeand).Two provenances of loblolly pine, with five open-pollinated families from each were analyzed fordifferences in height, volume, leaf area, and various crown characteristics. Families from theNorth Carolina and South Carolina Coastal Plain (ACP) and from the ?Lost-Pines? area of Texas(LPT) were included in the study. In addition to studying potential genetic variation, half of theplots received fertilization treatments so that potential nutrient and genotype x environment(GxE) interactions could be assessed. The stands were established in 1993 and height wasmeasured annually until year 3 when both height and diameter at breast height were measured.Leaf area measurements were made in 1999 using the LI-CORE LAI-2000 PCA and destructiveand non-destructive sampling of individual branches was done to estimate the crowncharacteristicsNutrient additions starting at stand establishment resulted in large gains in juvenile developmentin height, volume, leaf area, and growth efficiency. Fertilization also dramatically increasedfoliage and branch biomass at all crown levels, and also contributes to an early shift of foliagefrom the lower crown to the middle. This is most likely due to early canopy closure. The AtlanticCoastal Plain provenance consistently outperformed the Lost Pine Texas provenance in height,volume, and growth efficiency. The ACP provenance had more foliage, predominately in themiddle crown than the LPT provenance, which may explain some of the productivity differences.There was a significant amount of variation in height, volume, leaf area, and growth efficiencybetween the families of both provenances. Both the vertical distribution and quantity of foliage inthe ACP families may explain some of the variation in volume growth and growth efficiency,though neither show enough of a direct correlation to explain all of it. The crown characteristicsstudied seem to indicate that distribution of foliage itself is more important in explainingdifferences among the LPT families.<P>

<p>This study wasconducted in a 14-year-old midrotation loblolly pine (<i style='mso-bidi-font-style:normal'>Pinus taeda</i> L.) plantation to assess the effects of fertilizationand vegetation control (a) on soil net nitrogen mineralization, (b) litterdecomposition and (c) foliar nutrient concentrations and use, and vegetationgrowth. Fertilization (none, 224 kg haP) and vegetation control (none, complete) treatments were applied in a 2x2factorial design in March 1998. Soil net nitrogen mineralization was assessedby monthly field and laboratory incubations, litter decomposition and nutrient(N, P K, Ca, Mg, S, Mn, Zn, B, Cu) release dynamics were determined by thelitterbag method, foliar nutrient use of pines and hardwoods was estimated fromlitterfall and foliar nutrient concentrations, and growth was determined basedon annual measurements of diameter and height. Field net N mineralization rateswere 19, 18, 31, and 78 kg ha for control (C, notreatment), fertilization (F), vegetation control (VC), and vegetation controlplus fertilization (VC+F) treatments, respectively. Relative treatments responseswere similar in the laboratory incubations. Litter mass loss duringdecomposition was reduced by 9% after 32 months by vegetation control, andfertilization had no effect. The mobility of the nutrients was as follows: Cu≤N≤S<P<Zn≤Ca<K≤Mn<Mg≤B.Pine foliar N and P concentrations, and N and P use were significantlyincreased by fertilization, while effects of vegetation control on theseparameters were usually not significant. Fertilization significantly increasedannual volume growth of loblolly pine by up to 7 myr (32%) over the three years, while vegetation control had nosignificant effect.<P>

<p>MONTES, CRISTIAN RODRIGO. A Silvicultural Decision Support System for Loblolly Pine Plantations. (Under the direction of Dr. Lee Allen).Different heuristic techniques were used to solve the stand management problem. The techniques, together with a growth and yield model and an expert system, form the methodological components of a decision support system developed for deciding forest management activities throughout a stand rotation. Random search (RS), simulated annealing (SA) and tabu-search (TS), were used to solve the optimization problem. A Growth and Yield model for Loblolly pine in the Carolina piedmont was used to predict forest growth under different thinning and/or fertilization regimes. Silvicultural activities like competing vegetation control, soil preparation, and early fertilization were included as early growth treatments for different soil and site conditions. Response to establishment treatment was predicted using a knowledge system based on expected gain and growth response associated with several soil types. Sensitivity analysis was undertaken to fine tune solutions from SA and TS. Recommended parameters for SA were: alpha = 0.99, max-iteration = 100000 and min-temperature = 0.01. The combination of two diversification techniques: least-used move and swap move were found to interact for improving solutions using TS. Tabu-tenure was found optimal at 50 for a neighborhood size of 156. Tabu-Search proved to be the best technique of all.<P>

<p>The Hofmann Forest is a self-sustaining forest that provides the North Carolina State University College of Natural Resources with support for research, education, and extension service. The management of the Hofmann Forest requires data concerning historical records, complete and current resource inventory, and the ability to model future forest conditions. A geographic information system (GIS) database was created for the Hofmann Forest in 1992 to facilitate achievement of these data objectives. The database was not maintained or used regularly. Ongoing forestry research and silvicultural activities are constantly changing the resource conditions on the forest. This research examined a practical and accurate method for maintaining currency in the GIS database. Digital imagery was integrated into the original GIS database, and silvicultural records were used to update the existing data layers. Digital orthophotography, in the form of USGS Digital Orthophoto Quarter-Quads (DOQQs), was the primary source of imagery, but where the imagery was unavailable or contained insufficient spatial detail, unrectified aerial photographs were scanned, registered, and substituted. For the vegetation data layer of the GIS, spatial and attribute updates were completed and evaluated for silvicultural operations covering over three thousand acres. Some updates involved only changes in attributes. Spatial updates were completed with the digital orthophotography or digital aerial photographs; of these, some updates involved fairly simple spatial editing and others involved more complex spatial editing. The updates required the digital aerial photographs were all spatially complex edits. Acreage estimates accompanied the silvicultural records. GIS-derived area measurements were compared with those on the silvicultural records. There was not a significant difference between the two measures of area, however some discrepancies were present. A series of comparison tests were designed and performed to identify the potential elements of the area discrepancies. Spatial complexity of the editing procedure, different sources of digital imagery, and size of updated vegetation polygons were all examined. Degree of spatial complexity in the updates did not significantly contribute to area discrepancies. There was no significant difference in area discrepancies when either the DOQQs or digital aerial photographs were used. Size of the updated vegetation polygons was significantly negatively correlated with the discrepancies, showing that small absolute differences in area in small polygons result in large relative discrepancy values. Differentially corrected global positioning system (GPS) data were used to assess the horizontal positional accuracy of the GIS data layers. Following National Map Accuracy Standard (NMAS) guidelines, a sample of 25 ?well-defined? locations were collected using a Trimble GPS Pathfinder ProXR receiver with real-time differential correction capabilities. These same locations were identified on the Roads layer of the GIS database, the DOQQs, and the digital aerial photography. Root mean-square error (RMSE) was calculated for each of the different data layers, using the GPS data as reference locations. Only the DOQQ-derived points met the NMAS Class 2 horizontal positional accuracy standard. RMSE for the aerial photography and Roads layer were greater than the limiting RMSE for the NMSE Class 3 standard. Based on these results, it can be concluded that DOQQs possess greater horizontal accuracy than the digital aerial photography and are the preferred imagery source for the on-screen digitizing. Should greater resolution be required for a database update, orthorectification of the digital aerial photography could be used to correct horizontal positional errors. Recently, software packages have become available to orthorectify aerial photographs effectively and affordably.The presence of extraneous features in the vegetation layer of the GIS database almost certainly contributes to the area discrepancies. Features such as windrows, fire ponds, and logging decks are included in vegetation polygon area but not silvicultural record area estimates. Future database improvements should consider subtracting these features (and their associated areas) from the vegetation layer and creating separate database layers for each type of feature. A methodology report was developed to accompany the GIS database as a reference for future updating. Continuous maintenance of the Hofmann Forest GIS database is necessary to provide timely information for on-site forest managers and research activities, and to preserve an account of forest conditions that may be useful in present and future management decisions. On-screen digitizing with integrated digital imagery proved to be a feasible method for updating and maintaining the Hofmann Forest GIS database.<P>

<p> Collaborative forest restoration can reduce conflicts over natural resource management and improve ecosystem function after decades of degradation. Scientific evidence helps collaborative groups avoid undesirable outcomes as they define goals, assess current conditions, design restoration treatments, and monitor change over time. Ecological research cannot settle value disputes inherent to collaborative dialogue, but discussions are enriched by locally relevant information on pressing natural resource issues. I worked closely with the Uncompahgre Partnership, a collaborative group of managers, stakeholders, and researchers in southwestern Colorado, to develop research questions, gather data, and interpret findings in the context of forest restoration. Specifically, my dissertation (1) explored ways to better align collaborative goals with ecological realities of dynamic and unpredictable ecosystems; (2) defined undesirable conditions for fire behavior based on modeling output, published literature, and collaborative discussions about values at risk; (3) assessed the degree to which restoration treatments are moving forests away from undesirable conditions (e.g., homogenous and dense forests with scarce open habitat for grasses, forbs, and shrubs); and (4) looked at the validity of rapid assessment approaches for estimating natural range of variability in frequent-fire forests. </p><p> The current practice of defining desired future conditions pulls managers and stakeholders into command-and-control thinking and causes them to dream away resource tradeoffs and the unpredictability of forest change. Instead, moving ecosystems away from undesirable states and reducing unacceptable risk might allow for diverse and socially acceptable conditions across forested landscapes. The concept of undesirable conditions helped the Uncompahgre Partnership come to agreement over types of fire behavior and stand conditions they wanted to avoid through management. I determined that restoration treatments on the Uncompahgre Plateau are generally moving forests away from undesirably dense conditions that were uncommon prior to Euro-American settlement. My assessment was largely based on data collected during collaborative workdays with the Uncompahgre Partnership. Our rapid assessment approach for estimating historical forest structure took a quarter of the time required for scientifically rigorous stand reconstructions, and it provided reasonably accurate estimates of tree density and spatial patterns. </p><p> Our data on historical stand structure revealed that fragmentation and loss of open grass-forb-shrub habitat between tree groups were the most dramatic and undesirable changes occurring in frequent-fire forests over the past century. Many restoration treatments are focused on restoring spatial patterns in tree groups, with little attention to spatial patterns in open grass-forb-shrub habitat. I determined that the juxtaposition of tree groups with grass-forb-shrub habitat >6 m from overstory trees is important for restoring understory cover, diversity, and composition. Focusing on undesirable conditions in stands, such as high tree density and scarcity of grass-forb-shrub habitat, can help collaborative groups find common ground and design treatments that restore structure, composition, and processes in forest ecosystems.</p>

Diss. (sammanfattning) Umeå : Sveriges lantbruksuniv., 2001. / Härtill 5 uppsatser.

Diss. (sammanfattning) Umeå : Sveriges lantbruksuniv., 2001. / Härtill 5 uppsatser.

<p> Improved equations for predicting future dominant height, diameters, and number of surviving trees in a forest stand were developed for loblolly pine in the mid-Gulf region of southern United States using tree data from 115 stands across the region. The data were split into two sets and models were fitted on each data set using contemporary statistical modeling approaches in SAS^&reg; and R^&reg; software. Several models were fitted and compared. Fitted models were evaluated based on two-fold cross validation techniques. The best equations had high fit indices and acceptable prediction standard errors. Model parameter estimates were significant at 5% significance level and exhibited logical model behavior. In the future, the system level performance of these equations will be evaluated after which the equations will be incorporated into the Cutover Sites Loblolly Pine growth and yield simulator developed and maintained by the Mississippi Forest and Wildlife Research Center.</p><p>

Abstract not available